ANALYSIS AND INTERPRETATION OF THE POTTERY

58 BUREAU OF AMERICAN ETHNOLOGY

^[Bull.173

ordered sequence. In the ceramichorizon, drills, polished

and pecked

celts, chloritic schist

and

clay pipes, fragments

from

soapstone

vessels,

and

natural quartz crystals are

most common.

In the pre*

ceramic sites, crude axes,

end

scrapers,

and

the hafted variety of

end

scrapers are found in greater numbers.

What

this

means

^is that if

these artifacts are found in sufficient

numbers

at a site they

may be

useful in establishing the general position of the site in the ceramic or preceramic time sequence for the area.

From

another point of

view

these artifacts of infrequent occurrence can serve as

an

excellent check

on

the other data

from

a site.

For

example, if a site

seems

to

conform

to

one

part of the time sequence but has a series ofcelts, pipes, drills, etc.,

which seem

to fit

more

closely to the opposite part of the sequence, the data suggest that possibly the site

had been

occupied

by two

groups at widely different times

and some method must be

derived to separate the artifacts into

two

distinct groups.

For

this purpose of serving as a double check, the less

abundant

miscellaneous artifacts

were

always considered in this study.

No*57^P"^^^*

CULTURAL PATTERNS,

VIRGINIA

— ^{^HOLLAND} 59

Albemarle Pottery Series (Evans, 1955, pis. 4, 5, 6, fig. 3):

A

group of pottery types on a ware typically light red to orange, sometimes gray-red, sandy textured with a crushed-quartz temper (rarely with crushed granite or greenstone) which are angular,

medium

to large particles. Manu-

factured by coiling. Decoration consists of punctures with a sharp stick or

narrow slits on the rim. In the majority of cases the rim is fairly vertical or tapersslightly inwardandonly rarelyslopesoutward. Theshapesare typically round-bodied ^pots with straight sides or with a slightly constricted collar and

short vertical rim.

The above-mentioned ware has been classified into five pottery types based on surface treatment: Albemarle Plain, Albemarle Cord Marked, Albemarle Fabric Impressed, Albemarle Net Impressed, Albemarle Simple Stamped, and Rivanna ^{Scraped. Albemarle} Cord Marked surfaces had been beaten with a cord-wrapped paddle, the cord being, in most cases, a simple, double- twisted, two-strand cord ranging from

medium

to coarse in size. The surface

was paddled when the clay was moderately soft. The fabric impressions had beenmade with a plain plaited fabricwith close, fine weft and a

medium

coarse to wide, heavy warp. Albermarle Net Impressed variety had been marked with a wide open knotted net deeply impressed in wet clay.

Albemarle Simple Stamped had been beaten with a paddle wrapped either with smooth thongs or roots or a paddle with faint grooves. Rivanna Scraped variety was produced by an irregular tool when the clay was very wet.

Marcey Creek Pottery Series (Evans, 1955, ^pi. 12, ^fig. 6):

A

^{group of pottery} types characterized by a light-tan to red-brown to gray- red color, soft paste, soapy texture and feel, crushed-steatite temper, very irregular, uneven, lumpy ^surfaces. The majority of vessels were apparently hand modeled, patched, or kneaded, while a few sherds suggest coiling. The

rims are fairly thin compared to the body wall and are either vertical or out- sloping. There is an occasional nicked rim by way of decoration. The sherd samples suggest direct copy of typical steatite vessels which are either oval or rectanguloid bowls with flat bases, irregular surfaces, curved to straight sides

with an occasional handleat the ends.

Two

types are recognized: Marcey Creek Plain and Selden Island Cord Marked. The plain type is smoothed by hand only, rough to the feel, very un- even and irregular with lumpsof temperprotruding through the paste. Selden Island Cord Marked, impressed with a cord-wrapped paddle, has a haphazard, overlapping, crisscrossing or diagonal pattern.

New

River Pottery Series (Evans, 1955, ^pi. 13, fig. 7):

A

^{group of pottery types on a ware} characterized by a gray-tan ^surface, with crushed-shell temper, incompletely fired in an oxido-reducing atmosphere pro- ducing a gray-cored paste. The majority of the sherds suggest modeling or patching as the method of manufacture. Decorations, which are frequent on the rim sherds, are gashes, finger pinchings along the lip, lower edge of the folded-over rim or along the collar. There are rounded loop handles and ^gen-

erally the shapes are round jars with globular bodies, short to medium-sized necks, the orifice smaller than the greatest body diameter and with a recurved or vertical rim.

Four types have been defined.

New

River Knot Roughened and Net Im- pressed has been paddled or rubbed on the exteriors with a knotted net, leaving a coarse, rough surface with impressions of knots and a few of the mesh lines.

Usually the mesh of the net is obliterated, suggesting roughening with a crum- pled net.

New

^River Cord marked type has been haphazardly beaten with a

60 BUREAU OF AMERICAN ETHNOLOGY

^[Bull.173 cord-wrapped paddle forming parallel or crisscrossing patterns. Nonoverlap- ping, parallel impressions are most common.

New

River Fabric Impressed type was treated with a fabric of plain-plaited or twisted varieties. Often the fabric has been applied several times in one area as if wrapped on a paddle or around the hand.

New

River Plain has both interior and exterior surfaces

smoothed over, but is still fairly uneven and irregular; sometimes the interior is scraped.

Radford Pottery Series (Evans, 1955, pis. 16, 17, ^fig. 9):

A

group of pottery types on a ware characterized by gray to gray-tan color,

a gray to black core resulting from incomplete firing in an oxido-reducing fire,

crushed-limestone temper, and with distinctive rim and vessel shapes. The

rims are either incurving, straight, or incurving with a thickened or folded-over

lip, decorated withfingerpinching, smallgashes,lightlyincised lines and, rarely, araised rib. Thereare infrequentstrap handles. They were generallymade by hand modeling or patching, although some sherds show coiling. The vessels arerounded jars with globular bodies, the orifices of which are smaller than the bodies.

Thisserieshas fourpottery types. Thesurfaces of RadfordKnot Roughened and Net Impressed had been beaten with either a net-covered hand or paddle, creating a haphazard, overlapping, rough surface with knot and cord impres- sions. It was apparently treated when the clay was leather dry. Radford Cord Marked type had been beaten with a cord-wrapped paddle without too

much

overlapping or crisscrossing. Radford Fabric Impressed sherds are im- pressed with a plain-plaited or twined fabric. The majority suggest that this fabric was wrappedaround apaddleor the hand and the exterior surfacebeaten or rubbed. Radford Plain Type issmooth butuneven.

Stony Creek Pottery Series (Evans, 1955, ^pis. 18, 19, 20, fig. 10):

A

group ofpottery types characterized by fine quartzsand temper, gritty and sandy texture, light tan to light orange or light red-tan surfaces, fired in an oxidizing atmosphere. Coiling as the methodof manufacture is evident on the majority ofthe sherds. Typically, there is no decoration.

A

^{variety of} shapes were reconstructed from the sherds. There were deep open bowls with sub- conoidal to conoidal bases; globular-bodied jars with conoidal bases and with insloping straight sides forming an orifice smaller than the body diameter, and

tallpot forms with conoidalor subconoidalbases.

Seven types are recognized. Stony Creek Cord Marked has been treated with fine to medium-fine cords, typically in a crisscrossing, well-executed, over- lapping pattern, usually diagonal to the rim. Stony Creek Fabric Impressed type is difficult to analyze because the sandy nature of the paste caused the sherd surfaces to erode easily.

A

characteristic ^of the type is the faintness of the fabric impressions even on uneroded surfaces, suggesting application when

the clay was leather dry. Stony Creek Simple Stamped sherds have been beatenwithapaddlewrappedwithsmooththongsorthin, smoothroots or grass.

The decorated type, Nottoway Incised, has the exterior treated as the fabric- impressed type, then the incisions were

made

with a flat, blunt stick, applied crudely andunevenlyto therimand body. Motifsare diagonallines extending from the lip, haphazard crosshatching, double chevrons, paired lines, and tri- angles. Stony Creek ^Plain has smoothed, fairly even and regular interior and

exterior surfaces. Rivanna Scraped is scraped or combed with a tool, leaving small irregular serrations.

The

basic differences in temper, paste, fii'ing, manufactm*e,

and

decoration of each pottery type are fah-ly easy to distinguish once

one

No.^StT"

^'^^'

CULTURAL PATTERNS,

VIRGU^IA

HOLLAND ₆₁

has

had

a ^{little experience} with pottery typology

and

classification.

The

numerical

and

^percentage

breakdown by

series for each sherd col- lection arranged

by

^{site is found in} the Appendix:, table 7.

The

per- centages were plotted as bars

on

strips of

graph

paper with a scale of

1 cm. equal to 10 percent arranged in

columns

wide

enough

to

accom- modate

the greatest percentage in

any

one series.

A

^collection of 50 or

more

sherds

was

considered usable with a fair degree of accuracy (Ford

and WiUey,

1949, pp. 35-37); these

were

plotted as solid bars.

The

four collections with sherd samples

between

22

and

47 were plotted with diagonal lines to suggest uncertainty of the results

due

to a small sample.

The bottom

of the seriation ^(fig. 8)

was

fixed, in part,

by

referenceto

Evans' ceramic study (1955, ^fig. 18).

He had

noted potter}- types characteristic of the Central

and North

Central

Ceramic

Ai-ea spilling

over into the northern ^district of the Allegheny

Ceramic

Ai-ea (ibid.,

pp. 103, 108). Therefore, reference to his seriation in the former area indicated that sites with a high percentage of

Stony Creek

Pottery

Series, relatively smaller percentages of Albemarle Series,

and

the presence of

Marcey Creek

Series are in the lowest part of the sequence.

The

bar graphs of

two

sites,

AU-26 and AU-11,

having the three requu-ements

mentioned above

were placed at the

bottom

of the seria- tion (fig. 8). According to Evans' study the

Stony Creek

Series

would

fade as the Albemarle Series blossomed ^(ibid., p. 100, fig. 18); thenext four sites

(AU-13, RM-4, AU-31, and AU-21)

were arranged accord- ingly.

Good

trends,

though

foreshortened

by

the limited

number

of sites, appeared.

The Stony Creek

Series decreased

from

60 percent to 10 percent, while Albemarle increased

from

32.2 percent to 89 percent.

Following

on

the sLx-site sequence developed above, sLx

more

sites

(AU-45, EB-3, AU-5, AU-35-V-1, AU-44, and AU-35-V-2)

with Albemarle

and Stony Creek

Series

were

seriated to continue

and

de- velop the best trends of the first sLx sites (fig. 8). It

became immedi-

ately apparent that a

new

ceramic influence

had

reached the area.

The Stony Creek

Pottery Series, as expected, continued to decline to 3.2 percent

and

2.7 percent, but the Albemarle Series, instead of con- tinuingto increase as it

had

in the Central

and North

Central

Ceramic

Area,

began

to fade as the

Radford

Pottery Series blossomed. This

new

influence, although present in lovr percentages (up to 5.8 percent) at the

bottom

of the seriation increased to 40.3 percent at the top of the 12-site sequence.

The

five remaining sites with ceramic samples presented a problem.

Fom- had

high percentages of

Radford

Pottery ^Series, only one

had

a trace of Albemarle,

and

only one

had any Stony Creek

^Series.

Were

the samples

(HD-9 and RB-7)

foundin the

two museum

collections so 471762—60 5

62 BUREAU OF AMERICAN ETHNOLOGY

[Bull. 173

o

ex,

co

<u CO

Xi

iixX

1)o a 3

cr

GO

No*57l^"^^^"

CULTURAL PATTERNS,

VIRGINIA

— ^{^HOLLAND} 63

highly selected that they

would

not

show any

trends in a seriation study? Traces of the

New

^{River Series}

^{had been}

^{present in the 12-}

sitesequence

akeady

developed

and

there

was none

in the foursamples with a high percentage of

Radford

Series.

The

fifth site

had

a high percentage ^of the

New

River Pottery Series

and

a very low percentage of

Radford

Pottery Series.

To

resolve these difficulties it

was

decided to seriate the four sites

(HD-2, RB-7, BA-1, and HD-9)

with the high percentages of the

Radford

Series as a

group

to continue the trends

which had been

de- veloped

by

the sequence (fig. 8).

HD-2, ^an

^{excavated midden,}

had

a trace (3.5 percent) of Albermarle Series

and

the lowest percentage of

Radford

Series (96.5 percent). This

was

placed

on

the

bottom

of this four-site sequence,

and

the others fell into place with increasing per- centages of

Radford

Series

and an

absence of Albermarle. Obviously, then, a

gap

existed in the sequence. Alore sites

would

doubtlessly bridge the

gap by showing

a progressive loss of Albermarle and

an

increasing percentage of Radford.

Stony Creek

^Series pottery,

which had

already declined to a trace,

would

either continue as a trace (as

on RB-7)

or

would be

absent entirely.

The

probabihties are that the

New

RiverSeries

would

not appear in

any

large percentages

and more

than Hkely

would

be absent

on most

of the sites.

The

question of selectivity ofsherds in the

museum

samples

(HD-9

and RB-7)

is not

answered

positively.

However,

they seriate well with

two

excavated

middens

having 96.5 percent of

Radford

Series or more,

which

indicates that thsre

was

a blossoming of the

Radford

Series to such a degree that it represented practically the entire ceramic

complex on

certain sites in the survey area.

The

question

now

arose as to

whether

the top of the seriated se-

quence

was

represented

by

the blossoming of the

Radford

Series (fig.

8). This did not

seem

possible for

two

reasons. In the northern dis- trict of the Allegheny

Ceramic

Area, the

Keyser Farm

^site, proposed

by

Griffin ashavingbeen occupiedin the post-Columbian era

(Manson, MacCord, and

Griffin, 1944, p. 413),

had

a ceramic

complex

with variants of the

Radford and New

River Pottery Series.

With

^this temporal assignment

and

the trace of

New

^{River Series in sites so far}

discussed, it

was

expected that the top of the seriation

had

not been reached. In the southern ^{district of} the Allegheny

Ceramic Area

the excellent sequence developed

by Evans

(1955, ^fig. 19)

showed

that here

Radford

blossomed at the

bottom

of the sequence

and

faded with the expanding of the

New

River Series.

With

these

two

studies as guides, it

was

obvious that the top of the seriation sequence should be represented

by

relatively large percentages of the

New

River Series

and

either

no Radford

Series or

moderate

percentages ofit.

The

only

site to fit this distribution is

AU-2,

and, ^{therefore, it} has been placed

64 BUREAU OF AMERICAN ETHNOLOGY

[BuU. 173 at the top of the seriation (fig. 8).

Agam

the intervening sites

mth

Radford

Series pottery declining

and New

River Series ppttery in- creasing

have

not been discovered as yet in the survey area. This lack is

shown on

the seriation chart

by

a gap.

In

summary

the following generahzations

may

be

made

^:

The

earli- est ceramic complexes of the Central

and North

Central

Ceramic Area

existed coevally in the northern district of the Allegheny

Ceramic

Area.

Two

distinctive ceramic complexes, limestone-tempered

Rad-

ford Pottery Series

and

^shell-tempered

New

River Pottery Series,

moving

in

from

the west or southwest, disrupt the ceramic patterns of this latter district but

do

not

have any

influence

on

patterns in the Central

and North

Central

Ceramic

Area.

The

^earliest of these

new

ceramic influences, the

Radford

Series, appears in percentages of 96.5 to 100 percent at a relatively later period in the area of this survey than it does in the southern district of the Allegheny

Ceramic

Area.

In the southern ^district it appears as a well-developed

complex

at the

bottom

of the seriation for that area. In the survey area there is a gradual transition

from

the

Albemarle and Stony Creek

tradition to the 100 percent

Radford

tradition.

The

late ceramic influence, the

New

River Series, is not well represented in the survey area but its

occasional presence is of importance in

marking

the

most

recent time

levels.

Let us

compare

the results of the point

and

blade seriation (fig. 5)

and

the seriated sequence based

on

pottery (fig. 8),

which were

derived independently of each other. It

was

impossible to collect uniformly large samples of both pottery

and

chipped-stone materials

from

ce- ramic-bearing ^sites. Therefore certain sites appearing

on

the ceramic seriation are not found

on

the projectile-point

and

large-blade study,

and

vice versa, although a tabulation of the artifact

sample

is

shown

in tables 6

and

7 of the Appendix.

Although

there is not 100 percent

agreement

in the order of those sites that appear in

both

sequences, a sufiicient

number

are in the

same

relative positions to suggest that the

two

forms of evidence can be used independently with

some

degree of accuracy.

To

be specific, the relative positions of sites

AU-11 and

AU-13

are the

same —

^{at the}

^bottom

^{of the} pottery-type sequence

and

at the beginning of the ceramic-bearing sites in the point

and

^blade sequence.

However,

their sequential positions are reversed, a matter of not too serious concern at ^this stage of the comparison. Sites

AU-21, AU-45, AU-35-V-1, and AU-35-V-2 mamtain

not only ^their relative positions but also their sequential positions. It will

be

re- called that although the stone artifact

sample

of

RM-4 was

so

mixed

with

RM-4A

^{that it}

was

impossible to separate the various ^cultural complexes

and

use the

sample

successfully in the point

and

blade seriational studies, the pottery

sample was

clearly a unit in itself. Its

No*57]^"